There are two primary errors investors make. The first is to invest in an idea that doesn’t play out as expected, creating a loss. The second is to avoid making an investment in the first place that would have done well. One is an error of action; one is an error of inaction.

Both are huge problems. But let’s focus on inaction. Investors, as a percentage of the total population, peaked in 2000 at the height of the tech bubble. Either burned or out of money to invest in, the total numbers of investors has declined even as the overall population has increased.

Many folks sat out the market rally that started in March 2009. They’ve missed out on one of the longest enduring rallies of all time. Percentage-wise, the gains are sizeable as well.

But they may have a buying opportunity in the next 12-18 months. That’s because one of the most reliable indicators of financial trouble ahead is starting to form. It hasn’t yet—but once it does, it has a 100 percent track record of forecasting a recession. And while every recession is different, they do lead to a decline in the stock market.

What’s this indicator? The yield curve. That’s a key indicator in the credit market that there’s trouble ahead. Remember, the bond market is where cautious money goes to avoid the speculation of stocks, commodities, cryptocurrencies and everything else under the sun.

In the bond market, yields tend to be positive over time. If you’re locking money into a fixed-income product for 10 years, you want a higher yield than for a 2-year bond to account for all the various risks that can occur over a longer time period. And that 2-year bond should have a higher yield than a 3-month bond. This accounts for various risks that can occur over time, usually risks pertaining to changes in interest rates.

That’s why a typical yield curve, looking from maturities anywhere from 3 months to 30 years out, should be positive. The longer the time period, the bigger return an investor should get. Yield curves work best on an apples-to-apples comparison, so there are yield curves for U.S. Treasury bonds, junk bond yields, high-grade corporate credit, municipal bonds, and so on.

Let’s focus on U.S. Treasury yields, because things are starting to shift there. Specifically, the yield curve is starting to flatten out. That’s when the difference between time periods isn’t as strong. It means you might be better off buying a 10-year yield now vs. a 30-year, since you may be able to roll that into a higher yield when it comes due. You might not, and that’s a risk too.

In a worst-case scenario, the yield curve inverts. That’s when interest rates are higher on the shorter end than the longer end. In essence, the bond market starts pricing in sour economic data, and the possibility of lower interest rates, which tend to cause yields to fall in long-dated bonds. As the slow-moving, invested-for-safety capital in the Treasury market shifts, shorter-duration bonds start to look more attractive than longer-duration ones.

Yield curve inversions are rare, but occur with an unusual regularity. The last one appeared in 2007 on the eve of the financial crisis. The one before that appeared in 2000, at the height of the tech bubble that led to a massive drop in stocks greater than the economic decline that came with it. Nobody can say with certainty what the next recession will look like, but if the yield curve inverts, there’s a good chance of a major drop in stocks within a year of that forming.

The problem with this inversion is that it tends to occur during periods of greed in the markets. It’s like a canary in the coal mine that the workers have long stopped paying attention to. We shouldn’t.

It’s not time to throw in the towel on this mega-rally yet. The curve is flattening and may not even invert. We’ll likely get one last big hurrah that brings in the greed like 2007 or 2000 all over again.

But as one of the most surefire recession indicators, the shifting yield curve worth paying the most attention to. And, it’s always worthwhile to have a plan in place.

Typically, before a market crash becomes evident, there’s first a lot more volatility in stocks. We haven’t seen much volatility in the past few years, and 2017 has been historically quiet on an exceptional level. Only a handful of trading days have featured a decline anywhere near 1 percent.

That will be the first major change. If you can’t stomach volatility, it may be time to move into some slower-moving companies with a heavy focus on maintaining and growing a large dividend—like AT&T (T) as mentioned in last week’s blog.

Second, it may be worthwhile to take some profits off the table. That will raise some cash for future buys at lower prices. I’d wait for an actual inversion in the yield curve first. But there are other ways to raise cash, such as covered call writing on existing positions. Besides earning extra income, the strike price of a call option allows you to set a fixed price you’d be willing to sell at.

Finally, when the yield curve does invert, fear-based trades like gold or going long volatility should take off. So should a strategy of investing a small amount of capital into long-dated put options against a market index. These strategies are designed to profit from the market’s decline, but also allow you to stay invested on the long side.

After all, one of the biggest challenges is staying in while times are tough. Avoiding pain is one thing, but many investors who sell out when things aren’t going well don’t get back in—and the error of inaction rears its ugly head.

Bottom line: The market is starting to throw its biggest yellow flag in a decade. It might not amount to anything. But it might. There’s plenty of time, and plenty of ways to prepare. Do so.

Andrew Packer is a Senior Financial Editor with Newsmax Media. He currently writes the Insider Hotline investment advisory, serves as investment director for the Financial Braintrust, and writes the monthly newsletter Crisis Point Investor.